We have included the case studies guide in the help in Version 0.1.
What will we do concerning CAST and the ILRI sets in R-Instat. In the case studies guide we state that it would be good if statistics training could be more problem-based. That fits well with the ILRI set of case studies. It is also a project that would be done again - except it already exists. But I suggest it is hardly used. Can we do anything about this in R-Instat. Can we "make it easy"? Should we do this now?
One aspect is that the ILRI resources are reasonably easy to find, and we could check with Jane Poole. They are here:
From here you can download the whole resource. If you don't, i.e. if you use it on-line, then I don't see how you can get easy access to the data sets that are used. It should be relatively easy to put them into our File > Open from Library dialogue. Not trivial, but relatively easy for most of them. I suggest this could be worthwhile. (These are all based in Africa.)
A second area is the access to sensible textbooks. CAST is now with a new lease of life, given the recent changes Doug has made. And we have special books for e-sms and e-siac. When people use CAST they often ask about the data sets, because they would like to do their own analyses. They confuse CAST's role as a textbook with a statistics package. Could we make at least some of the data sets available in R-Instat? Is it worth the time? How do we link to CAST sensibly anyway? Do we want to? I suggest the least I would like is a Resources menu item in the help, with a page (or more) on CAST. Agreed? We could also have a video?
On data sets here is what Doug says:
"Regarding the data sets, I’m afraid that I don’t have them in an external format (other than a few that I managed to find in Excel format to start with). Basically, when I needed a data set for any particular situation, I usually hunted around in books, journals and web sites and just typed the values straight into HTML. It isn’t too hard to extract the values for any data set you want from their textual format in the HTML. Each variable is stored in a “param” tag with its values space-delimited. You could copy each such text string into Word, globally replace all spaces with carriage returns, then paste into an Excel column.
Regarding extra context, the Data Sets link on the left in CAST gives references to where I found the data (if that is known). That’s all the extra information that I have for most."
David is also interested in large open data sets. I would welcome a strategy overall.
We have included the case studies guide in the help in Version 0.1.
What will we do concerning CAST and the ILRI sets in R-Instat. In the case studies guide we state that it would be good if statistics training could be more problem-based. That fits well with the ILRI set of case studies. It is also a project that would be done again - except it already exists. But I suggest it is hardly used. Can we do anything about this in R-Instat. Can we "make it easy"? Should we do this now?
One aspect is that the ILRI resources are reasonably easy to find, and we could check with Jane Poole. They are here:
https://www.ilri.org/biometrics/default.htm
From here you can download the whole resource. If you don't, i.e. if you use it on-line, then I don't see how you can get easy access to the data sets that are used. It should be relatively easy to put them into our File > Open from Library dialogue. Not trivial, but relatively easy for most of them. I suggest this could be worthwhile. (These are all based in Africa.)
A second area is the access to sensible textbooks. CAST is now with a new lease of life, given the recent changes Doug has made. And we have special books for e-sms and e-siac. When people use CAST they often ask about the data sets, because they would like to do their own analyses. They confuse CAST's role as a textbook with a statistics package. Could we make at least some of the data sets available in R-Instat? Is it worth the time? How do we link to CAST sensibly anyway? Do we want to? I suggest the least I would like is a Resources menu item in the help, with a page (or more) on CAST. Agreed? We could also have a video?
On data sets here is what Doug says:
"Regarding the data sets, I’m afraid that I don’t have them in an external format (other than a few that I managed to find in Excel format to start with). Basically, when I needed a data set for any particular situation, I usually hunted around in books, journals and web sites and just typed the values straight into HTML. It isn’t too hard to extract the values for any data set you want from their textual format in the HTML. Each variable is stored in a “param” tag with its values space-delimited. You could copy each such text string into Word, globally replace all spaces with carriage returns, then paste into an Excel column.
Regarding extra context, the Data Sets link on the left in CAST gives references to where I found the data (if that is known). That’s all the extra information that I have for most."
David is also interested in large open data sets. I would welcome a strategy overall.